SUR with different sample

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

ccheong
Posts: 2
Joined: Mon Jan 15, 2018 5:38 pm

SUR with different sample

Hi All,
I am using Eviews10. Beginner user. Learning how to use the Seemingly Unrelated Regression (SUR).

Is it possible have different sample within a system (SUR)?
Example:
y = c(1) + c(2)*x {sample 1}
y = c(3) + c(4)*x {sample 2}

The main purpose is to compare C(2) against C(4). i.e. C(2)-C(4)=0

Kind regards,
Chee

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 11923
Joined: Tue Sep 16, 2008 5:38 pm

Re: SUR with different sample

No.

Of course you can always lag your variables.

startz
Non-normality and collinearity are NOT problems!
Posts: 3405
Joined: Wed Sep 17, 2008 2:25 pm

Re: SUR with different sample

Copy x into x1 for the first sample, leaving NA elsewhere. Then the same for x into x2 for the second sample.

Note that all that's really going on is a fancy way of doing least squares on each equation.

ccheong
Posts: 2
Joined: Mon Jan 15, 2018 5:38 pm

Re: SUR with different sample

Hi Startz,

Sample 1 doesn't overlap with sample 2.

y = c(1) + c(2)*x {sample 1: large firms}
y = c(3) + c(4)*x {sample 2: small firms}

Kind regards,
Chee

startz
Non-normality and collinearity are NOT problems!
Posts: 3405
Joined: Wed Sep 17, 2008 2:25 pm

Re: SUR with different sample

smpl 1 100
series x1 = x
smpl 101 200
series x2 = x
smpl @all

y = c(1) + c(2)*x1
y = c(3) + c(4)*x2